Skip to main content
Upcoming release — Zylon 1.58These docs are published ahead of release so you can prepare your API integrations. The release image is not yet available. For installations running 1.57, use the 1.57 documentation.
ZylonGPT is the model-facing API. Most apps follow a simple loop: send messages, attach or reference artifacts, generate embeddings for search, and let the assistant call tools when it needs structured data or retrieval.

Authentication and scopes

Use a gateway-scoped token from Developer Console. PrivateGPT.Read.All allows the endpoints below. An account token with Read.All reaches the same allowlist only when it carries gateway context through gateway_id; Read.All can be created without gateway_id, but that token cannot call ZylonGPT.
  • POST /api/gpt/v1/messages
  • POST /api/gpt/v1/complete
  • POST /api/gpt/v1/messages/count_tokens
  • POST /api/gpt/v1/messages/validate
  • POST /api/gpt/v1/embeddings
  • GET and HEAD requests under /api/gpt/v1/skills
  • POST /api/gpt/v1/skills/validate
Async message lifecycle routes, every artifact route, skill mutations, and unmatched fallback routes require a read-write scope and return 403 for a read-only token. Use PrivateGPT.ReadWrite.All for a gateway-scoped token or ReadWrite.All where account-token access is appropriate. Multiple scopes combine their grants, but effective permissions remain limited by the token owner’s roles. See Token Management for creation defaults and validation rules.

Messages

Send synchronous messages with read access; async lifecycle operations require read-write access.

Artifacts

Ingest, retrieve, list, or delete collection content with read-write access.

Embeddings

Turn text into vectors for search and similarity.

Tools API

Enable semantic search, tabular analysis, web fetch, and custom tools.